[thune]

     Quote:

http://www.benchmarkmedia.com/discuss/feedback/newsletter/2011/12/2/0-ohm-headphone-amplifier-sonic-advantages-low-impedance-headphone-amp

"All measurements shown in this paper are taken at the headphone inputs."
 
I honestly don't care what the voltage waveform distortion is at the headphone terminals with a high impedance source, I care what the acoustic distortion is from the transducer. The plots in that article give zero idea what the transducer distortion will look like. If they showed the current waveform distortion from their 0-ohm source, it would look ugly too. A dynamic driver is ultimately a current driven device, and there are potential reductions in distortion by using a high impedance source.
 
Of course, there are the obvious changes in frequency response from higher impedance sources, which are real issues. But to me, this article is completely misleads regarding actual transducer distortion in sources with non-zero impedance: it says nothing.
 

 

[kiteki]

 
Hmmm.
 
Does Beyer offer an adapter for the DT48? How does that work?
 
I heard that headphone is all machined aluminium and has special drivers and it's really expensive to make or smt.

 

[SanjiWatsuki]

 
Quote:

Indeed, not 100% THD+N, but 0.028%, but the reason why THD + N decreases with the increase of the damping fact is still a mystery to me, do you have a hypothesis?
 

 
I've got two ideas off the top of my head. One, resistors add noise over a circuit, if I recall, so it will inherently have a higher THD + N. The other possibility is that it has to do with the actual damping on the driver and that a more tight control over the driver is able to, by a small amount, have less THD. 

 

[SanjiWatsuki]

Honestly, output impedance doesn't really matter to a DT48-E, at least. Their impedance is very uniform across the entire frequency spectrum, so there are no deviations in frequency response. Tyll's measurements showed an increase in THD + N when a resistor was added to the output of the amp, so I really think it is perfectly fine and arguably better to run them without the 120 ohm resistor in line, even if they were designed for it. 
 
There might be some business with the bass that would go on, but it's already a very fast headphone. I really don't think that the speed of the bass would change that much with an inline resistor. 
 
Quote:

 
Hmmm.
 
Does Beyer offer an adapter for the DT48? How does that work?
 
I heard that headphone is all machined aluminium and has special drivers and it's really expensive to make or smt.

 

 

[kiteki]

 
What does THD+N matter? The Etymotic ER-4PT sounds a lot better when you add a 75 ohm resistor.

 

[khaos974]

Quote:

Quote:

Indeed, not 100% THD+N, but 0.028%, but the reason why THD + N decreases with the increase of the damping fact is still a mystery to me, do you have a hypothesis?
 

 
I've got two ideas off the top of my head. One, resistors add noise over a circuit, if I recall, so it will inherently have a higher THD + N. The other possibility is that it has to do with the actual damping on the driver and that a more tight control over the driver is able to, by a small amount, have less THD. 

It seems to be THD considering the harmonics we see in the FFT graphs.
 
 

 

[SanjiWatsuki]

It really doesn't unless it gets excessive. In this case, I really don't believe the change in THD+N is audible. For an example of THD+N being able to make a difference, look at InnerFidelity's graph of the THD+N for the MDR-V6. The rapidly increasing noise in the sub-bass region says to me that the driver is struggling to make those low frequencies, and will have a degree of distortion in the sound if I equalize that region up. In that case, I would care about the THD+N. Generally speaking, though, if it stays below 1% for most frequencies on the headphone and below like 3% for the super sub-bass frequencies, I don't care about it.
 
As for the Etymotic, I believe that resistor creates an upper treble focus, like I said earlier for how resistors affect single BA IEMs. To calculate it at 8khz, we'd so something like:
(27) / (27+75) = .264
(40) / (40+75) = .347
20 * log(.347/.264) = 2.37dB shift.
 
Basically, the ER-4PT would upwards shift the upper treble frequencies, slightly more and more as the frequency goes up,  with the adapter cable... or to quote Etymotic themselves: "[size=small] [/size][size=small]restores some high frequency response normally delivered by ER-4S"[/size]
 
It's a fancy way of saying "we're emphasizing the upper treble."
Quote:

 
What does THD+N matter? The Etymotic ER-4PT sounds a lot better when you add a 75 ohm resistor.

 

 

[anetode]

 
Quote:

 

 
Oh my! Now, I take issue with this graph. A damping factor of 2 is NOT going to give you 100% THD+N. Do you know how bad 100% THD+N would be? How unbearable that would be? I'm sorry, but that's definitely a measuring artifact of some sort, there is NO WAY that you'd get that kind of distortion out of this. If a damping factor of 2 resulted in that much distortion at peaks, you can bet that there would be a ton of moaning about odd awful sounding artifacts from high impedance sources.
 

This misunderstanding appears to have been cleared up, though just to make sure: are you referring to the fundamental harmonic's -0db peak as 100% THD?
 
On the rest I agree with your analysis: any source-related differences an order of magnitude lower than the audibility threshold for THD are simply not going to matter.
 
That's not to say that going out of one's way to add specific types of distortion to the signal can't be fun

 
Quote:

 
What does THD+N matter? The Etymotic ER-4PT sounds a lot better when you add a 75 ohm resistor.

My taste tends to the ER-4P rather than the ER-4S sound sig, but it's brave of etymotic to point out that the R is the only difference between the two. I can see the value of a reasonably priced and flat IEM in the 4S, especially for music producers. With such results from one measly armature, I wonder why everyone gets worked up about shoving eight of those into a custom shell. It goes counter to the whole crossovers as a necessary evil doctrine.

 

[kiteki]

 
The ER-4PT/S is quite technically capable, but 3/4/6 driver custom IEM's are more immersive and joyful.

 

[SanjiWatsuki]

Quote:

 

This misunderstanding appears to have been cleared up, though just to make sure: are you referring to the fundamental harmonic's -0db peak as 100% THD?
 
On the rest I agree with your analysis: any source-related differences an order of magnitude lower than the audibility threshold for THD are simply not going to matter.
 
That's not to say that going out of one's way to add specific types of distortion to the signal can't be fun

 

 

I tried to add a little bit at the end to clarify, but I was a little vague. Gist of what I was saying is just that you wouldn't expect to legitimately hear the headphone have some abnormal 106dB artifact at certain frequencies during normal listening at 106dB.
 
 

 

[Chris J]

Quote:

 
Then why is everyone up in arms now about the measured negative effects?

Kiteki:
Virtually all headphone amplifiers are designed as voltage sources.
Strict electrical theory states that an ideal voltage source has zero output impedance.
The headphone amp applies the voltage across the headphone impedance, the amp supplies current proportional to how much impedance the cans have.
So, in theory, you are trying to apply the same voltage waveform to the headphone as the amp receieved at it's input. 


 
Quote:

I put the resistors in a dongle cable makes it very easy. 

N-God:
I guess I know what I'm doing when I get some spare time!
 
 
Quote:

Whether to run on a high or low output impedance amp has little to do with what the listener likes (OK, maybe a little) but more to do with how the headphones were originally designed. 
 
There is no easy & reliable way to predict how a given speaker will react to being driven by an amp with an output impedance other than what it was designed for. The safest assumption is that the effects will happen around the crossover points, but the crossover points could come up at a totally different frequency when you change the output impedance of the amp!  Your blanket statement of what frequencies changing the output impedance of a speaker amp will effect is totally inaccurate. 
 
Running headphones designed around a 120ohm output impedance on a 120 ohm output solves all of these problems with no work from the end user.
 
The biggest problem with the conflicting 0ohm and 120 ohm standards is that the only MFR that even comes close to stating that they follow one or the other is Beyerdynamic. Everyone else just leaves the user on his own figuring it out. I'd trust people to use their own judgement of what sounds good and go from there so why dont we get together and do that? 

Regarding running headphones off the output impedance they were designed for.........................if only ALL headphone manufacturer's would clearly state: "this headphone is designed to work with this output impedance".
 
Regarding multi-driver loudspeaker systems:
I would think that any multi-driver loudspeaker system would be designed to work best around a solid state amp of virtually zero output impedance. Can you think of any manuf. that doesn't? I would also guess that any high end multi-driver loudspeaker manuf. would also try their speaker out on a tube amp, but only to ensure it still basically sounds OK.
 
 


Quote:

That's a mighty, mighty sneaky way of presenting the information! Showing the benchmark in negative dB is a very underhanded way of showing the data. Let's put those in a bit more balanced numbers.
 
 

S/W:
Showing noise and distortion below 0 dB is pretty standard in this type of graph:
For example:
It allows you to look at the graph and read the siganl to noise ratio off it very easily.


Quote:

     Quote:

"All measurements shown in this paper are taken at the headphone inputs."
 
I honestly don't care what the voltage waveform distortion is at the headphone terminals with a high impedance source, I care what the acoustic distortion is from the transducer. The plots in that article give zero idea what the transducer distortion will look like. If they showed the current waveform distortion from their 0-ohm source, it would look ugly too. A dynamic driver is ultimately a current driven device, and there are potential reductions in distortion by using a high impedance source.
 
 

Thune:
Like I said earlier in my reply,
The headphone amp applies a voltage across the headphone voice coil.
Current is then drawn proportional (actually inversely proprotional) to how much impedance the can has.
You are correct in that current is what creates the magnetic field on the voice coil which is what causes it to move in the magnet structure.
Ideally you want the voltage applied across the voice coil to be an exact replica of what came out of the DAC and into the headphone amp. 
 
Quote:

 

 
I've got two ideas off the top of my head. One, resistors add noise over a circuit, if I recall, so it will inherently have a higher THD + N. The other possibility is that it has to do with the actual damping on the driver and that a more tight control over the driver is able to, by a small amount, have less THD. 

Hi SanjiWatsuki:
The noise created by a 120 ohm resistor (or any value less than that) would be negligible and can be ignored at the output of a can amp. So would the distortion created by the actual resistor, the distortion we are seeing is created by the interaction between the amp output impedance and the headphone impedance
I suspect that the added distortion we are seeing in the Benchmark article is actually created by the reduced damping factor caused by the higher output impedance, i.e.the amp cannot control the headphone driver motion as well as it could at a lower (or virtually zero) output impedance. So we are seeing the back EMF created by the poorly controlled headphone driver.
 
Folks, try out high and low output impedance yourself

.
See what you prefer.
Me, I like my solid state can amp output impedance to be LOW.